Papers

Molecular dynamics simulations are used to compute the potential of mean force (PMF) between two united-atom methane molecules at several temperatures and integration time-steps. At a fixed time-step, the contact minimum of the PMF deepens with increasing temperature, as expected for hydrophobicity driven association. Compared with a time-step of 0.5 fs, one that preserves equipartition, larger time-steps alter the PMF and make the contact minimum more favorable. Thus even relative free energy values are sensitive to time-steps that break equipartition. Using quasichemical theory, we partition the free energy of association into hydrophobic and hydrophilic contributions. The hydrophilic contribution opposes association and is insensitive to the time-step. Thus the artificial enhancement of association at larger time-steps comes entirely from the hydrophobic contribution. We explain this behavior through the temperature dependence of the internal pressure of the liquid. The same analysis also accounts for earlier observations of the liquid's p-V behavior under conditions that break equipartition.

Within the geometric optics approximation, the optical expansion scalar describes the local rate of change in the amplitude of a wave or the area of a light bundle propagating through the Universe. As such, it is fundamental for describing almost all the observations we make in cosmology. It was recently pointed out that this quantity can, in principle, be reconstructed from measurements of the angular diameter distance $D_A$ and the effective observed expansion rate $\mathfrak{H}$. Motivated by this possibility, we introduce the observable $\vartheta \equiv \theta/\omega_o$, where $\theta$ is Sachs' optical expansion scalar, and $\omega_o$ the observed photon frequency. We then derive the linear perturbative expression and angular power spectrum for $\vartheta$ in a flat FLRW universe with scalar perturbations. Comparing the power spectra with those of the fluctuations in $D_A$ and with power spectra from relativistic N-body simulations reveals no immediate striking advantage of studying the fluctuations of $\vartheta$ to those of $D_A$. However, the anisotropies of $\vartheta$ do exhibit several unique qualities such as an enhanced sensitivity to the peculiar velocity and relativistic potentials. We suggest that our results may hint at $\vartheta$ being useful for developing consistency tests of cosmological observations and possibly probing departures from FLRW cosmology and modified gravity theories. The formalism developed and presented here provides the foundation for future studies into these possibilities.

Smartphone cameras and computer vision (CV) hold significant promise in assisting public agencies with enforcing traffic laws and enhancing road safety. This work designs and tests a smartphone-based method for automated speed estimation and vehicle identification (license plate, make/model, and color recognition) via an automated pipeline to assist enforcement agencies in reliably identifying speeders. The CV code accurately recognizes nearly half (46%) of the license plates' text on 1,800 images from a Brazil open-source dataset, called UFPR-ALPR. Code tests on daytime recordings from hand-held smartphone videos (n = 73) and roadside cameras (n = 42) in Austin, Texas yield 60.8% accuracy for color detection (among all possible RGB color categories), 48.6% on vehicle make/manufacturer identification, and 16.89% on vehicle make and model identification. Prediction accuracy for speed estimation (within a 20% range), vehicle make (within the top 3 predictions), and license plate recognition (within the top 10 predictions) are 16.3%, 16.9%, and 29.7%, respectively. This paper also illuminates the legal, technological, and practical aspects of using smartphones for enforcement, including the potential use of recordings for enforcement purposes, emphasizing the need to transform the potential of smartphone-based CV technologies into practical tools for vital information on traffic violations.

Statistical model transformations based on randomly stopped sums, maxima and minima are widely used to extend statistical models. We characterize the complete set of stopping models for which randomly stopped sum and extreme model transformations function as statistically stable (idempotent) model extensions. Stability requires the underlying stopping model to be closed under pgf composition. We prove that any finite-dimensional, connected stopping model closed under pgf composition is necessarily a family of random variables whose pgfs commute. Using the corresponding Koenigs function, we establish that these models form a statistical manifold admitting a global, one-dimensional parametrization $\theta = \Pr(N=1) \in (0, \theta_*]$, where the probability mass at $i$ is a polynomial in $\theta$ of degree at most $i$. Finally, we establish a duality between stopping models closed and containing the identity variable (the ones yielding stable extensions) and the set of probability distributions supported on the positive integers. These findings disprove the long standing conjecture that statistical stability occurs only under geometric stopping.

We prove that for every isotropic log-concave distribution $P$ on $\mathbb{R}^d$ and every even $m\ge2$, the polynomial $(Cm)^m\|v\|_2^m - \mathbb{E}_{X\sim P}\langle X,v\rangle^m$ is a sum of squares, where $C>0$ is a universal constant. This improves on the Poincaré-dependent bounds (Kothari and Steinhardt, 2017), recovering the optimal moment bounds for log-concave distributions. As an immediate corollary, we obtain computationally efficient algorithms with dimension-free error guarantees for a wide range of statistical estimation problems. Our proof proceeds by using stochastic localization to decompose $P$ as an average of strongly log-concave measures, whose centered moments admit the subgaussian certificates (Diakonikolas et al., 2025). With a covariance-adapted choice of localization, we show that a fourth-moment certificate derived from the variance inequality for quadratic forms (Letwin, 2026) suffices to control this averaging at every even degree.

High Performance Computing (HPC) and Quantum Computing (QC) systems are increasingly converging towards unified High Performance Computing-Quantum Computing (HPCQC) infrastructures, driven by a growing need to bridge classical and quantum workflows, which affects all levels of the system stack, from the hardware to compilers and runtimes, all the way to applications. However, today's QC devices are still in the Noisy Intermediate-Scale Quantum (NISQ) era, are error-prone and resource-limited, and therefore require specialized optimizations and topology mappings to achieve sufficient fidelity. This places special emphasis on proper compilation and optimization within the overall quantum software stack. Many existing stacks remain fragmented, with separate components responsible for device selection, compiler-pass optimization, and job queue scheduling. This paper proposes a unified, learning-based selector that integrates these disparate stages into a cohesive framework. Our proposed selector scheme leverages reinforcement learning and deep learning models that can be extended to simultaneously optimize multiple objectives – such as fidelity, compilation time, and scheduling latency – while dynamically adapting to circuit characteristics and device conditions.

This paper establishes that transcendental entire functions do not have wandering Fatou components whose forward orbit is bounded. This settles a long-standing question in transcendental dynamics. A stronger statement is proved using the same method: the orbit of every point in a wandering domain of a transcendental meromorphic function is unbounded. In particular, transcendental meromorphic functions do not admit orbitally bounded wandering domains. The proofs rely mostly on uniform estimates for hyperbolic area, following the approach of Ye [arXiv:2609.23834v1].

Most ANN-to-SNN conversion methods rely on a specific correspondence between the source activation and the spiking neuron dynamics. We propose a finite-state continuous-time Markov chain (CTMC) neuron framework whose stationary spike flux can approximate every continuous nonnegative monotone activation function on a compact interval. For a generalized CTMC family with affine input-dependent transitions, we prove uniform approximation to arbitrary accuracy over this function class and derive an explicit approximation error bound. In practice, two- and three-state CTMCs fit ReLU, sigmoid, softplus, and clipped ReLU on the evaluated input ranges, and we evaluate corresponding MLP conversions for each activation with layerwise rate scaling. Moderate clipping improves the conversion cost-accuracy tradeoff on the MNIST MLP and reduces SynOps by 27% on VGG-11/MNIST at matched ANN-SNN accuracy gap criteria, whereas the trend reverses on VGG-11/CIFAR-10. Mean-field and layerwise diagnostics indicate that finite-window sampling and terminal-layer mismatch are the main residual errors. Overall, our results establish finite-state CTMC neurons as a theoretically grounded framework for activation-flexible ANN-to-SNN conversion beyond fixed activation-neuron correspondences.

We introduce a class of tensor-network states in which physical legs are shared among local tensors, termed leg-tied tensor ansätze (LETTA). Physical leg ties encode long-range correlations directly, while a virtual matrix product state (MPS) backbone retains short-range multipartite entanglement. The linear virtual backbone allows us to develop a deterministic density matrix renormalization group-like variational optimization algorithm using exact contractions over the active tie-boundary sets and local minimization. We demonstrate the advantages of LETTA for the two-dimensional frustrated $J_1$–$J_2$ Heisenberg model and the three-dimensional transverse-field Ising model. Our results show that LETTA is substantially more accurate than same-bond-dimension MPS calculations and can typically reach the accuracy of much larger MPS calculations using one order of magnitude fewer variational parameters. LETTA thus opens the door for explicitly correlated tensor-network states that can encode long-range correlation beyond virtual bonds.

Recent advances in paraphrase detection reveal a fundamental trade-off: large language models achieve high accuracy but require high computation, while efficient Siamese-BERT variants offer practical scalability with reduced transparency in rationale generation. We present R-DEIM Net, a 76M-parameter dual-expert architecture exploring whether moderate-scale models can achieve competitive accuracy on paraphrase detection while enabling human-readable rationale generation. The architecture combines two specialized components: an Interaction Expert that captures token-level similarity patterns through multi-scale 2D convolutions and attention head allowing variable input length, and a Reasoning Expert that uses a Flan-T5-small decoder to generate rationales as auxiliary supervision. Rather than re-encoding generated text, we extract and pool decoder hidden states as complementary features for classification. On the Quora Question Pairs dataset, R-DEIM Net achieves 90.07% accuracy and 90.16% F1-score via 10-fold cross-validation. This represents competitive performance with strong transformer-based baselines (e.g., MFAE BERT: 90.54% accuracy) and recent large language model based approaches (LLaMA-70B) while using a substantially smaller parameter budget. The model generates rationales alongside predictions, providing potential for auxiliary human-readable descriptions.

Heterogeneous accelerator architectures offer an efficient path to performance for compute-intensive workloads. However, full-stack integration remains difficult. We present EAAC (Extensible Accelerator Architecture), a flexible and extensible compiler and hardware architecture designed to lower the overhead of hardware-compiler co-development for rapid prototyping of hardware accelerators. EAAC targets static data-flow workloads with predictable memory access patterns. By using MLIR, we enable possible integration with a range of different frontends that emit MLIR. And by providing a minimal set of compiler functionality that enable necessary data-orchestration we lower the effort needed to get a simple implementation of a hardware acceleration unit up and running, while providing a relatively blank and un-opinionated starting point for further work. We validate this approach with a prototype GEMM accelerator combining a systolic array unit with an embedded RISC-V core, compiled end-to-end through the EAAC MLIR pipeline. On a synthetic fully-connected-layer workload, the accelerator achieves a 28x speedup in execution time over a RISC-V-only baseline, with the GEMM operation itself accounting for only 1.64% of total execution time. We further characterize the compiler's hardware-semaphore allocation, showing that the number of semaphores required scales linearly with instruction count in the worst case, and identify this as a concrete target for future optimization.

We study the largest common subtree of two independent unlabeled uniform attachment trees (also known as random recursive trees). Our main result shows that, when the two trees have $n$ vertices each, their largest common subtree has at least $n^{0.83}$ vertices with high probability. This is obtained by starting with the common subtree induced by the Ulam–Harris labels in the two trees and improving using local optimization steps. We also give some upper bounds and bounds for general random tree growth models. We leave as an intriguing open question to understand the magnitude of the size of the largest common subtree.

We study Fibonacci polyominoes, a class of column-convex polyominoes whose lower boundary is a staircase path made of unit horizontal and vertical steps. We derive generating functions that enumerate these polyominoes by area and semiperimeter. The resulting formulas give refinements of the classical Fibonacci enumeration and lead to explicit expressions involving Catalan numbers. We give a bijection with labeled Dyck paths that provides combinatorial proofs of these formulas and translates natural statistics on Fibonacci polyominoes into peaks, returns, and related statistics on Dyck paths. This yields refinements by descents and left-to-right minima, including distributions governed by Narayana numbers. We also study consecutive columns of equal height, leading in the diagonal case to a refinement of the Catalan enumeration governed by Motzkin numbers.

Video diffusion is computationally expensive, as it requires executing a large model across many denoising steps. Even with step-distillation, inference remains expensive because every distilled step still requires a costly model evaluation. We present TRACK: TRajectory-Aware Capacity routing via top-K selection, a heterogeneous denoising strategy that switches between compatible large and small models at selected steps, reducing the average cost per denoising evaluation. The switching steps are determined using a calibration process. TRACK first rolls out a reference trajectory with the large model. Then at each step, the small model's prediction is also collected and compared against the large model's prediction to obtain a relative disagreement score. Both models receive the same latent, timestep, conditioning, and guidance inputs. Aggregating this signal over a calibration set produces a disagreement score map across diffusion steps, which determines a switching policy for an efficient inference process: quality-sensitive steps keep using the large model, while steps with low disagreement scores are routed to the small model. Inference executes only the selected model at each step, requiring no retraining, architecture or scheduler changes, or online dual-model evaluation. Across Wan 2.1, Cosmos 3, TurboDiffusion, and FastVideo, TRACK yields $1.95\times$, $2.04\times$-$2.73\times$, $2.69\times$, and $2.17\times$ speedups, respectively, with comparable aggregate quality and high diversity retention. TRACK thereby establishes automated, training-free model switching as a practical acceleration paradigm for video diffusion.

Cosmological-collider signals are usually interpreted as clean probes of massive particles present during inflation. In this work, we study a massive spectator field excited by a spatially quenched random source, which may arise from a frozen realization of a light field or environmental sector whose relaxation time is long compared with the inflationary timescale. We determine which features of the collider signal remain reliable probes of the heavy field. After Gaussian disorder averaging, the random source modifies only the statistical component of the propagator, leaving the spectral function and heavy-field poles unchanged. The particle spectrum is therefore preserved, and the logarithmic frequency of the collider oscillations remains fixed by the heavy-field mass. What changes is how the massive modes are excited and imprinted on cosmological correlators: the temporal profile of the source controls the amplitude, phase, and envelope of the oscillations. For a power-law source, the fixed-realization problem is exactly solvable in terms of Lommel functions. Decaying sources can make the nonanalytic clock signal dominant in the squeezed limit, while spatial white noise can destructively interfere with the standard collapsed-trispectrum signal. Thus, mass measurements remain robust against this class of quenched disorder, whereas amplitude-based inferences can depend sensitively on the random environment.

Large language models increasingly operate as persistent assistants in user-facing, shared-session, and tool-augmented settings. When users disclose sensitive information during an active conversation, that information may remain behaviorally recoverable through later prompts even after the dialogue shifts to unrelated topics. We introduce PrivDrift, a benchmark for auditing whether user-disclosed secrets remain recoverable after conversational topic drift and persuasion-based probing. PrivDrift contains 1,000 controlled multi-turn dialogues with seeded secrets, content-dense drift turns, and standardized extraction probes. Across three LLMs with extended context windows, dialogue-level hybrid leakage remains substantial, ranging from 38.7% to 54.6%, and varies strongly by model, secret type, and persuasion intensity. Within the tested drift window, additional topic drift does not reliably reduce leakage, suggesting that privacy risk in active LLM contexts should be evaluated as a persistent behavioral failure mode rather than only as training-data memorization or immediate jailbreak behavior.

Élie Cartan established that the metric and intrinsic curvature of a $D$ dimensional embedded manifold $M$ could be determined by tracing the response of another surface $N$ of the same dimensionality, as it is rolled without slipping and twisting on $M$. In the context of spacetime geometry, this construction underpins the MacDowell-Mansouri formulation of General Relativity. We consider extensions of this framework that correspond to rolling of a shape with twisting and with shape evolution; it is shown that these naturally describe torsion and non-metricity respectively. Cartan-geometric formulations of teleparallel gravity and symmetric teleparallelism are discussed in addition to further manifestations of non-metricity in first-order formulations of gravity.

While Vision-Language-Action (VLA) models demonstrate impressive capabilities in robotic manipulation, their memoryless nature renders them brittle to test-time environment shifts, particularly hardware shifts caused by wear or imperfect calibration. Enabling these models to self-adapt during deployment without requiring continuous on-site recalibration remains a critical bottleneck for real-world scalability. In this work, we introduce Self-Adaptive VLA, a novel post-training recipe that enables the policy to iteratively adapt to deployment-time hardware shifts leveraging its own rollouts as context. To do so, we first collect policy rollouts under deliberately injected hardware shifts. We then transform the base policy's training data into shift-conditioned expert demonstrations by pre-compensating the expert actions for these known shifts. Next, we introduce a lightweight, plug-in context encoder that compresses the context, including visual observation, proprioception, and actions in the shifted environment, into a latent context token. This token modulates the policy through adaptive layer normalization (AdaLN). Furthermore, we find that context tokens can be ensembled, allowing the policy to iteratively self-correct and mitigate failures step by step. Extensive experiments across four precision-critical bi-manual and dexterous manipulation tasks show that Self-Adaptive VLA recovers over 80% of the base policy's performance under hardware shifts, such as actuation bias and joint encoder offsets. Moreover, Self-Adaptive VLA enables more robust deployment to new workstations compared to the base policy. Our approach provides a pathway for robust large-scale real-world robot deployments and easier maintenance. See videos at https://icefoxzhx.github.io/self-adaptive-vla.

Let $R$ be a commutative ring and $\Phi$ an inverse family of ideals with greatest proper member $\mathcal{L}$. We introduce and study $\Phi$-reduced and $\Phi$-coreduced modules, extending the corresponding notions for powers of a single ideal. We show that, on these classes of modules, the generalized torsion and completion functors associated with $\Phi$ are determined by $\mathcal{L}$. We establish characterizations and closure properties of these modules and obtain a Greenlees-May type adjunction and a Matlis-Greenlees-May type characterization. When $\Phi$ is a system of ideals, we further investigate the radicality of the generalized torsion functor and derive associated torsion theories on suitable Serre subcategories. These results provide a module-theoretic framework for studying generalized torsion and completion with respect to families of ideals.

We establish global $W^{2,p}$ solvability for the Poisson equation with mixed Dirichlet–Neumann boundary conditions in two classes of domains in all dimensions $n\geq 2$. For $C^{1,\alpha}$ domains with a Reifenbeg flat interface, we obtain the optimal range $1<p<4/3$, provided that $\alpha>1-1/p$. For Lipschitz polyhedra with facewise boundary decompositions, we obtain solvability for $p$ close to $1$. In both settings, we establish endpoint $W^{2,1}$ solvability for data in an adapted atomic Hardy space. Applications of the methods to Lamé systems are also discussed.

Approximate pattern matching (APM) on highly repetitive texts is a central task in bioinformatics. Bi-directional r-indexes support left- and right-extension of a pattern and thereby accelerate APM algorithms based on search schemes, but existing variants – br-index and b-move – suffer from two bottlenecks: $O(\sigma)$ character-predecessor/-successor queries on the run-length-encoded BWT per extension, and predecessor queries on sparse bit vectors to maintain a value in the suffix array interval and to access the PLCP array while locating. We present Move-rb, a bi-directional r-index built on the optimized r-index Move-r. Although Move-rb is 2x larger than br-index, it is up to 24% smaller than b-move, answers APM queries 1-4 orders of magnitude faster than br-index, 1.9-10x faster than b-move and up to 5.5x faster than the state-of-the-art bi-directional (uncompressed) FM-index columba, which is 36-42x larger. Memory usage (including index size) during APM locate queries is reduced by 1.5x (up to 6.5x) for Hamming distance and 2.5x (up to 7.4x) for edit distance. Move-rb can be constructed 3-14x faster while using 19-141x less memory than br-index, b-move and columba. A variant using a relative Lempel-Ziv-encoded suffix array locates up to 10x faster while being 1.2-2.4x larger. We achieve these speedups by optimizing index operations and search scheme APM algorithms: Without a direction switch, Move-rb computes an all-$k$-character extension in output-optimal $O(k)$ time and a single-character extension in the same time. Augmenting Move-rb with $O((r+\overleftarrow{r})\log(n/r_{\min})\log\sigma)$ bits reduces a single-character extension to $O(\log\sigma)$ time, where $r$ and $\overleftarrow{r}$ are the numbers of runs in the BWT of the text and its reverse, and $r_{\min}=\min(r,\overleftarrow{r})$. A direction switch incurs only $O(\log\log_\omega(n/r_{\min}))$ additional time.

Graph-structured optimization with linear constraints is fundamental to critical infrastructure but faces scalability limits due to massive strict hard constraints and high dimensionality. While recent projection-based methods such as Trainable Sampling Kaczmarz-Motzkin Net (T-SKM-Net) guarantee feasibility, they face high computational costs in dynamic environments by processing the entire constraint set and requiring expensive matrix factorizations. To bridge this gap, we propose the Accelerated Trainable-SKM (AT-SKM) Net framework. To concentrate computation on the active constraints and eliminate redundant calculations, we introduce a hybrid sampling strategy guided by a topology-aware heterogeneous GNN model. To efficiently handle topological shifts in graph-based constraints, we employ a Cholesky Update mechanism that theoretically reduces the equality projection complexity from O(N^3) to O(N^2) under low-rank perturbations. Experiments on random geometric graphs, N-1 Security-Constrained DC-OPF, and minimum-cost gas transport problem demonstrate that AT-SKM reduces iteration counts by up to 85% and achieves 2.95x-7.29x SKM layer speedups, while maintaining zero constraint violations.

We consider the decision of whether to return an existing draft answer or revise it using retrieved evidence, as in answer-revision systems. Draft confidence estimates whether the current answer is correct, but the decision requires estimating the effect of a specified revision. For offline training and evaluation, we grade both the returned draft and its candidate revision under the same correctness judge, which makes repair, harm, and the gap to an oracle observable. We call this paired effect its recoverability, and we train policies to predict it before revision. On 25,870 held-out open-domain questions across three revision setups, a scorer trained on the paired outcome has greater area under the accuracy–revision-rate curve than a matched draft-correctness scorer in all nine Llama setup–seed fits, and gains 0.23–0.68 accuracy points on average at development-selected thresholds, a difference significant across training runs only for dense retrieval. The resulting policy improves on always revising and on average closes more than a third of the oracle gap, although it still applies 38–46% of the harmful revisions. When a draft-free standard-RAG answer is also available, however, choosing between the draft and that answer is stronger by about two points for Llama and four for OLMo, and adding candidate revision as a third option yields no significant gain. Recoverability describes one revision; its value as an available action also depends on the alternatives.

In this paper we define a novel product on polyominoes called the tensor product of polyominoes. Using a special case of this product which we name a dilation, we construct a new class of polyominoes called the dilated closed paths. This class is a ”thick” generalization of Cisto and Navarra's class of closed paths (arXiv:2006.13935). We prove the Zig-Zag Walk Conjecture for this new class thereby fully characterizing primality for its associated polyomino ideals.

In multi-target regression, correlated targets are often coupled through multi-output Gaussian processes with an intrinsic model of coregionalisation (GP-ICM), assuming that sharing statistical strength improves overall performance. In practice, the benefits are inconsistent. Across the settings studied, we find that the main benefit of coregionalisation is joint uncertainty quantification rather than point prediction. Raw target correlation does not predict when coupling helps; in the separable GP-ICM settings studied here, residual correlation, the cross-target dependence left unexplained by independent per-target predictors, is the strongest predictor of joint-uncertainty gains. We introduce a lightweight diagnostic, $D_{\rm logdet}=-\frac{1}{2}\log\det R_{\rm res}$, which represents the idealised joint negative log-likelihood (NLL) gain from modelling a full rather than diagonal residual covariance and is computable from independent GPs alone. Across a controlled synthetic study, 16 multi-target benchmarks, and frozen transformer and convolutional neural network representations for keypoint regression, point prediction remains largely unchanged ($\Delta R^2\approx 0$). In contrast, $D_{\rm logdet}$ strongly predicts observed ICM NLL improvements ($\rho_s=-0.83$, $p<0.001$), outperforming heuristics such as the feature-to-sample ratio. We also propose Residual-ICM, which preserves independent marginal variances while adding residual-correlation structure to the joint covariance. Residual-ICM achieves the best average joint NLL among the compared methods, while the diagnostic indicates when covariance coupling is likely to be useful. The diagnostic is specific to global Gaussian residual dependence, the structure captured by separable coregionalisation.

Macroscopic materials function is determined not merely by equilibrium structure, but by how carriers, phonons, fields, defects, interfaces, and collective order dynamically evolve after perturbation. Ultrafast electron microscopy (UEM) uniquely bridges this gap, coupling femtosecond-to-nanosecond timing with real-space, reciprocal-space, and energy-resolved contrast. Here, we review how these integrated capabilities now quantitatively map energy flow and conversion in electronic materials, decode electronic-structural coupling in quantum and correlated systems, and isolate the localized propagation of carriers, polaritons, strain, and phonons across optoelectronic and nanomechanical architectures. We emphasize the predictive design rules now emerging from direct metrology of carrier-lattice coupling, momentum-resolved phonon thermalization, defect-controlled phase transformations, and authentic operando switching. We critically assess UEM's ongoing maturation from a qualitative, proof-of-concept experiment into a rigorous, quantitative materials platform. This paradigm shift is actively driven by integrated multimodal detection, advanced energy-resolved spectroscopy, high-repetition-rate sources, multidimensional (4D and 5D) acquisition, and physically constrained forward modeling. Ultimately, the central opportunity lies in establishing a complete, causal link from initial excitation to localized energy flow, transient structure, and macroscopic device function. By prioritizing quantitative reproducibility and robust mechanistic interpretation, UEM is positioned to engineer materials in functional states that equilibrium thermodynamics simply cannot describe.

Open-vocabulary Sound Event Detection detects and temporally localizes acoustic events described by arbitrary text queries. Progress in this emerging field is hard to assess: recent methods report on disjoint task subsets under incompatible protocols without a benchmark spanning the acoustic domains and query types the task presents. We establish a comprehensive benchmark by assembling six temporally-annotated tasks: four with fixed class vocabularies over domestic, urban and mixed indoor/outdoor scenes, plus two free-text grounding tasks. We evaluate five recent methods on identical data and metrics under a label-space zero-shot criterion. Our benchmark demonstrates that no prior method is competitive across all six tasks. We then introduce COSED, which surpasses prior work on five out of six tasks while staying on par with the best method on the sixth, with margins of 12-33% on three of them. COSED is the only system in our comparison competitive on every task, and so generalizes across acoustic domains and query types better than prior work. We also provide a leave-one-out ablation study that isolates the sources of the performance benefits: scoping negatives to their corpus of origin (25.8%), combining closed- and open-world supervision (16.8%), and improving temporal processing (16.4%).

Robust dexterous grasping requires maintaining physical stability despite contacts interactively evolving across the entire hand. A precomputed force distribution can easily fail under object motion, modeling errors, or external disturbances. In this paper, we present a framework for real-time force regulation over dynamically changing whole-hand contacts. Our method geometrically estimates contacts across all hand links using a tracked object model and proprioception, without requiring tactile sensing at those contacts. It repeatedly recomputes the desired contact-force distribution subject to friction constraints, actuator limits, and an actuation-consistency constraint motivated by classical whole-limb force analysis. We integrate this force-regulation controller with reactive reaching, enabling the hand to acquire a grasp, maintain it under disturbances, and regrasp after losing the object. Simulation experiments without gravity demonstrate improved grasp retention over fixed-allocation and fingertip-only execution under controlled perturbations, while real-world experiments on a 27-DoF arm-hand system demonstrate grasp maintenance and recovery under human-applied disturbances as contacts evolve across the whole hand. Project page: https://sangminkim-99.github.io/reactive-grasp-whole-hand/

Let $S=K[x_1,\ldots,x_n]$ be a polynomial ring over a field $K$ and let $I$ be a monomial ideal of $S$. We classify linear systems of parameters of $S/I$ over any field $K$ using linear algebra and show explicit linear systems of parameters when $K$ has at least $n$ elements. If $I(G)$ is the edge ideal of a perfect graph $G$, a cycle or the complement of a cycle, we show that $S/I(G)$ has a 0-1 linear system of parameters, and for graphs with independence number equal to $2$, we characterize when $S/I(G)$ has a 0-1 linear system of parameters.

Foundation segmenters such as SAM return several plausible masks for an unlabeled image, and a student trained on the wrong one inherits its errors. Choosing among them means querying a second large model or fitting a quality head to annotated masks. We show that a candidate can be judged by what it does to a frozen self-supervised backbone's features. Normalized DINOv2 patch features lie on a hypersphere, and a candidate mask splits that sphere in two. Based on this reading, we introduce SphereTrust, which scores each candidate by three properties of the split, the angular contrast between the two sides, the coverage of the foreground's appearance modes, and contact with the image frame, one for each of three common ways a mask fails, and ranks a pool in 0.55 s per image from the frozen features alone. On eight SAM and SAM3 candidate pools spanning camouflaged, salient, and dichotomous segmentation and camouflage under low light, SphereTrust exceeds the strongest evaluated external baseline on six pools by 1.7 to 9.3 percentage points in mean selected Dice. These comparisons include published selection rules and explicitly labeled adaptations of DSS and UCOD-MKD. On the two prompted camouflage pools, its mean selected Dice is within 0.1 percentage points of the candidate-derived DSS adaptation, with a lower catastrophic-error rate. Which cue carries the signal depends on the candidate pool. The same sphere also supports training. The leading candidates enter as a candidate set with their scores as priors, prototypes reorder them, and a cross-fitted second round completes the labels, raising weighted F by 4.5, 2.3, and 5.5 points over fixed-label training on the three MLLM anchor pools, with students competitive with published unsupervised methods on nineteen test sets.